Search for: All records

Many concurrent programs assign priorities to threads to improve responsiveness. When used in conjunction with synchronization mechanisms such as mutexes and condition variables, however, priorities can lead to priority inversions, in which high-priority threads are delayed by low-priority ones. Priority inversions in the use of mutexes are easily handled using dynamic techniques such as priority inheritance, but priority inversions in the use of condition variables are not well-studied and dynamic techniques are not suitable. In this work, we use a combination of static and dynamic techniques to prevent priority inversion in code that uses mutexes and condition variables. A type system ensures that condition variables are used safely, even while dynamic techniques change thread priorities at runtime to eliminate priority inversions in the use of mutexes. We prove the soundness of our system, using a model of priority inversions based on cost models for parallel programs. To show that the type system is practical to implement, we encode it within the type systems of Rust and C++, and show that the restrictions are not overly burdensome by writing sizeable case studies using these encodings, including porting the Memcached object server to use our C++ implementation.

Abstract This paper introduces and evaluates a general construct for trading off accuracy and overall execution duration in classification-based machine perception problems—namely, the generalized IDK classifier cascade . The aim is to select the optimal sequence of classifiers required to minimize the expected (i.e. average) execution duration needed to achieve successful classification, subject to a constraint on quality, and optionally a latency constraint on the worst-case execution duration. An IDK classifier is a software component that attempts to categorize each input provided to it into one of a fixed set of classes, returning “I Don’t Know” (IDK) if it is unable to do so with the required level of confidence. An ensemble of several different IDK classifiers may be available for the same classification problem, offering different trade-offs between effectiveness (i.e. the probability of successful classification) and timeliness (i.e. execution duration). A model for representing such characteristics is defined, and a method is proposed for determining the values of the model parameters for a given ensemble of IDK classifiers. Optimal algorithms are developed for sequentially ordering IDK classifiers into an IDK cascade, such that the expected duration to successfully classify an input is minimized, optionally subject to a latency constraint on the worst-casemore »overall execution duration of the IDK cascade. The entire methodology is applied to two real-world case studies. In contrast to prior work, the methodology developed in this paper caters for arbitrary dependences between the probabilities of successful classification for different IDK classifiers. Effective practical solutions are developed considering both single and multiple processors.« less

Abstract Reconciling the constraint of guaranteeing to always meet deadlines with the optimization objective of reducing waste of computing capacity lies at the heart of a large body of research on real-time systems. Most approaches to doing so require the application designer to specify a deeper characterization of the workload (and perhaps extensive profiling of its run-time behavior), which then enables shaping the resource assignment to the application. In practice, such approaches are weak as they load the designer with the heavy duty of a detailed workload characterization. We seek approaches for reducing the waste of computing resources for recurrent real-time workloads in the absence of such additional characterization, by monitoring the minimal information that needs to be observable about the run-time behavior of a real-time system: its response time. We propose two resource control strategies to assign resources: one based on binary-exponential search and the other, on principles of control. Both approaches are compared against the clairvoyant scenario in which the average/typical behavior is known. Via an extensive simulation, we show that both techniques are useful approaches to reducing resource computation while meeting hard deadlines.